Abstract
We have spent many years studying what has been happening to the American family. Haskins first addressed the issue when he was with the House Ways and Means Committee, which has jurisdiction over several social programs, especially the programs for adoption and foster care and the Aid to Families with Dependent Children program, which required knowledge of research on family well-being. He began writing about family issues shortly after he joined the Brookings Institution in 2001 (Haskins and Sawhill 2003). Sawhill’s first book, coauthored with Heather Ross, and published in 1975, was about the growth of single-parent families and their consequences for children (Ross and Sawhill 1975).
Like so many others in the scholarly and political worlds, we have been greatly influenced by the ideas of Daniel Patrick Moynihan. Now as the fortunate recipients of a prize given in his name, we happily acknowledge our indebtedness to his thinking, and we dedicate this article to his memory without in any way implying that he would have agreed with everything we say.
The history of his writing on the topic is well known. In 1965, as an assistant secretary in the Department of Labor, he wrote an internal report for members of the Johnson administration on his insights about the major impediment to continuing black progress (Moynihan 1965a). No paper written by a federal official has ever had as much influence as the Moynihan Report, as it was famously dubbed. The primary barrier to black progress in his view was weaknesses in the black family. He spoke of a “tangle of pathology” in the ghetto and was later ostracized both for his use of such language and for singling out the large proportion of single-parent families within the black community as an impediment to progress and thereby appearing to “blame the victim.”
In their excellent review of this period and its aftermath in an earlier issue of The ANNALS, Douglas S. Massey and Robert J. Sampson (2009), as well as William Julius Wilson (2009), all note the unfairness of these attacks on Moynihan. In particular, his critics at the time often ignored the fact that he believed the ultimate cause of the breakdown of black families was centuries of oppression and persecution, and that he argued for an affirmative effort to compensate for the consequences of that oppression, including a federal jobs program.
Nonetheless, fierce criticism of the Moynihan Report took its toll, not just on Moynihan personally but also on social science research and debate for decades (Rainwater and Yancey 1967). It would be a long time before any scholar or public intellectual would dare to address the sensitive issue of how differences in family structure affect school performance, employment, crime, and the related differences between blacks and whites. A few courageous souls broke the impasse by researching the way in which family structure affected both black communities and the lives of children, whether black or white, growing up in fatherless families. William Julius Wilson (1987), Sara McLanahan (McLanahan and Sandefur 1997), and those who contributed to the earlier issue of The ANNALS come to mind as first-rate examples of some of this newer scholarship.
Contributing to a more open discussion of these questions were data showing that the growth of single-parent families was increasingly affecting not just black communities but white and Latino communities as well. Charles Murray, in his book Coming Apart (2012), for example, focuses only on white families, placing declines in marriage at the center of his story of communities beleaguered by low levels of education, male joblessness, crime, and single-parent families. Other researchers, such as Andrew Cherlin (2014), note that as technological change and globalization have dried up high-paying jobs for those with little education, the white working class family has also begun to disintegrate. By 2000, the fraction (22 percent) of white single parents was as high as the fraction of black families that had so alarmed Moynihan in 1965 (Sawhill 2013). As Moynihan wrote in Family and Nation (based on his Godkin lectures at Harvard in 1985): “Certain jobs simply are no longer there, while the people who once held them are” (Moynihan 1986).
In his later writing, Moynihan evinced considerable humility about how much public policy alone could change this trajectory and the corresponding need for a more limited government than many of his liberal contemporaries thought desirable. He believed in the Catholic doctrine of subsidiarity and in the power of the smallest unit, the family, to shape lives. Although this view is often associated with conservative thinkers, Moynihan was no conservative (Weiner 2016). Not only did he support many liberal policies, from welfare to jobs programs, but he was the architect of the Family Assistance Plan, introduced by President Nixon shortly after his reelection in 1969. That plan would have replaced most other antipoverty programs, encouraged work, and provided a guaranteed income to low-income Americans. Its defeat in the Senate Finance Committee was due, in part, to the opposition of progressive advocates who believed that it was insufficiently generous despite the fact that by today’s standards it seems almost utopian. It would have provided a basic income of $1,600 to every American family with dependent children, or the equivalent of almost $10,000 in today’s dollars (Hess 2014).
Our conclusions on the family will be at least somewhat Moynihanesque. Like him, we think families are important and that marriage is beneficial to children. Our book, Creating an Opportunity Society (2009), and several issues of the Future of Children that we have coedited (McLanahan and Sawhill 2015; McLanahan, Donahue, and Haskins 2005) lay out a plethora of evidence that two parents are better than one. Indeed, we think there is now a consensus within the social science community about the adverse consequences for children of the breakdown of the family. And like Moynihan, we think a little humility is in order about what we really know about the causes of this breakdown and what, if anything, might work to restore the family. Finally, Moynihan believed that culture and other influences, such as economics and politics, were intertwined in a complex causal web, and we agree. As he noted, “the central conservative truth is that it is culture, not politics, that determines the success of a society. The central liberal truth is that it is politics that can change a culture and save it from itself” (Moynihan 1986). 1
Although much is owed to Moynihan, we address many developments that he could not have foreseen in 1965. One development is a growing class divide in marriage rates. In the 1950s and 1960s, marriage was not only widespread but varied little by education or other measures of socioeconomic status. Now marriage rates are linked much more strongly to measures of class, especially women’s education level. This raises the issue of the relative importance of class versus race in creating these trends.
A second development is the much greater prevalence of cohabitation—the new “marriage-lite”—and the challenges this creates for data analysis on living arrangements and for public policy, which is still based primarily on more traditional definitions of who shares living expenses with whom. Given that about 40 percent of American children spend some time in families headed by cohabiting parents by the age of 12, a new field of study focused on cohabiting parents has been established (Manning 2015).
A third development is new research suggesting that boys are more adversely affected than girls by growing up without a father, raising the specter of a vicious circle in which growing up in a single-parent family leads to diminished prospects for boys and young men, which then leads to even further declines in marriage.
A fourth development is the reform of welfare, something that Moynihan worked on during his long Senate career, even though he strongly opposed both the Clinton administration’s welfare reform bill and the bill that eventually passed in the mid-1990s. He condemned both in very strong language. 2 Cash welfare has been partially replaced by the Earned Income Tax Credit and other work supports that have fewer putative adverse effects on marriage and work than the old welfare system. But government taxes and benefits still penalize marriage in some instances, leading to calls for their reform.
A fifth development is the sharply rising educational attainment, especially of women, along with the associated tendency to delay marriage to later ages. A corollary to collapsing marriage rates is a rise in pregnancies and births among unmarried young adults in their twenties over the past three or four decades. Young people in their twenties may be delaying marriage, or foregoing it altogether, but they are not refraining from sex. Although teen pregnancy rates have declined sharply, the problem has moved up the age scale. Most of these pregnancies among the unmarried result from young single adults (under 30) drifting into unstable relationships. A report from the Guttmacher Institute found that 73 percent of pregnancies to unmarried women aged 20–24 were unplanned (Zolna and Lindberg 2012). These births also make women less marriageable, since most men they might later meet are at least somewhat reluctant to take responsibility for someone else’s child (Lichter and Graefe 2001).
In the remainder of this article, we briefly review the most recent evidence of trends in the family—as Moynihan famously said, we are all entitled to our own opinions but not to our own facts (Hess 2014)—and then explore two more complicated questions: the causes and consequences of these trends. And finally, we examine the various ways in which either government or nongovernmental institutions might slow or reverse these trends. In each case, we focus especially on recent research that has opened up new ways of thinking about these family issues and what as a society we ought to do in response.
Everyone’s Entitled to Their Own Opinions but Not to Their Own Facts
As shown in Figure 1, the growth of single-parent families has been dramatic (Cancian and Haskins 2014). From 7 percent of all families with children in 1950, the share of single-parent families increased sharply to 31 percent in 2015. Most of these single parents are mothers, but a moderate and growing proportion are fathers (22 percent in 2015).
Living Arrangements for Families with Children Under 18, 1950–2015
Another way of looking at the same trends is from the perspective of children. What proportion of children is born outside marriage? Again, the trends have been dramatic, although they have leveled off for most groups in recent years (see Figure 2). By 2014, about 29 percent of non-Hispanic white children, 53 percent of Hispanic children, 71 percent of black children, and 40 percent of all children in the United States were born outside marriage. It is the growth of unwed childbearing that has been the driver of increasing single parenthood since the 1980s. Divorce rates have actually declined slightly, especially among the well-educated.
Births to Unmarried Women by Race, 1970–2014
Children born outside marriage are not necessarily living in a single-parent family. An increasing number live with two cohabiting adults (see Figure 3). 3 In some cases, this includes both of their biological parents, but in just as many cases it includes a biological parent (usually the mother) and an unrelated boyfriend (Manning 2015). Children who live with two stable, committed biological parents appear to fare as well as those in married-parent families, but such durable relationships are relatively rare. Cohabitations tend to be fragile, especially those that occur as the result of an unplanned pregnancy. Nearly 40 percent of cohabiting couples who have a baby are no longer together by the time the child reaches age five, about three times the breakup rate for married couples over the same period (Hymowitz et al. 2013). This instability is not good for children (Manning 2015). Further, those in cohabiting relationships do not have the same legal rights as those in married-parent families. 4 If the couple were to marry, “marriage bonuses” or “marriage penalties” in tax or benefit programs would be created, often both. For example, eligibility for health benefits under Medicaid or the Affordable Care Act depends on the tax filing unit and thus excludes the income of an unmarried partner. By contrast, in the Food Stamp program (Supplemental Nutrition Assistance Program), the income of everyone in the household is counted. (We will say more about these marriage penalties and bonuses in the policy section.)
Families with Children Under 18 in Cohabiting Households
Despite the growth of single parents and unwed births among whites, the racial gaps are still wide and cannot be entirely explained by differences in socioeconomic status as measured by, say, educational attainment (see Figure 4). Of course, educational attainment is not a sufficient indicator of class, and a more robust analysis that included measures of wealth, employment, incarceration, and early death among black men, not to mention the adverse effects of residential and social segregation, would undoubtedly explain more of the racial gap in marriage rates. Still, most analysts have not been able to fully account for the differences (Raley, Sweeney, and Wondra 2015). While many people assume that slavery, which often broke up families, is at least partly to blame, the large black-white gap did not emerge until about the 1960s. 5 It is quite possible that the initial divergence in the 1960s was driven by some combination of limited job prospects for men and the availability of welfare for women, enabling them to support their children outside marriage. We say “enabling” rather than “causing” because evidence that welfare benefits caused the breakdown of the family is quite limited (Ellwood and Jencks 2004). But whatever factors caused a gap to emerge, culture and history are likely to play a role in sustaining it. For example, marriage rates are higher among Hispanics than among non-Hispanic blacks even though the former are economically just as disadvantaged (Raley, Sweeney, and Wondra 2015).
Births to Unmarried Women by Race and Education Level, 2014
Causes of the Trends
Most of the research that has attempted to explain these trends in family composition and living arrangements has focused on three primary drivers: increased opportunities for women; declining economic prospects among men, especially the least skilled; and changes in social norms or cultural attitudes. Although there is little consensus on the relative importance of each driver, most experts have concluded that both economics and culture have played a role (Sawhill 2014; Cherlin 2014).
The traditional model of marriage was based on the advantages of specialization within the household. Men, it was argued, had a comparative advantage in being the breadwinners while women were more suited to being homemakers (Becker 1974). But with the large-scale entry of women into the labor force, this model has become increasingly obsolete. The economic foundation of marriage is rapidly crumbling. Instead, marriage is now a symbol of commitment to another adult and facilitates joint parenting when children are involved. It also serves as a vehicle for the pooling of resources and for the mutual enjoyment of the lifestyle those resources make possible (Stevenson and Wolfers 2007; Lundberg and Pollak 1996; Reeves 2014).
Although the traditional model is on the wane, the perception that men should be breadwinners has not disappeared (Wang and Parker 2014). Women and men both still expect men to bring home a paycheck, and stagnant or falling real earnings for less educated men are widely hypothesized to have had an adverse effect on marriage. The empirical research on the size of this effect is quite mixed but tends to show that declines in men’s earnings have had modest effects overall (Sawhill and Venator 2015). A recent paper by David Autor, David Dorn, and Gordon Hanson (2016), using a sophisticated methodology to tease out the exogenous effects of trade-induced declines in employment opportunities among men, finds that the declines have had a significant but small effect on marriage rates. Several other studies have found that the effects of male employment on marriage are larger for blacks than for whites, and this can be traced to high rates of incarceration and very high rates of joblessness among young black men. In short, the evidence that marriage is disappearing among less-skilled whites is limited, but there is a shortage of marriageable men within the black community (Wilson and Neckerman 1986; Thomas and Sawhill 2002; Sawhill and Venator 2015). One promising response to this shortage is criminal justice reform.
The literature on the causes of the retreat from marriage tends to make two assumptions that are unlikely to hold true in the future. The first is the assumption that men will continue to be the primary breadwinners in most families. Although this remains true for now—and is a clear motivation for marriage—its influence is likely to erode over time. Even now, 40 percent of all primary breadwinners are women; and Edin and Nelson (2013) note that many men in low-income communities expect the mother of their children to support the family. A second assumption is that marriages will continue to take place within same-race groups. Currently, a very small fraction of all marriages are interracial, but this is changing rapidly (Frey 2014).
Accompanying these changes in the economic prospects of women and men, and interacting with them, has been a radical shift in cultural norms. Although most Americans still think that single parenting is a problem, they increasingly see marriage and childbearing as separate events and no longer disapprove of premarital sex, cohabitation, or births outside of marriage. The data in Figure 5 show how much more permissive attitudes toward premarital sex have become over the past four decades. An even more compelling illustration of shifting norms may be the way in which language has evolved. Forty or fifty years ago, a woman who had had sex before marriage was called “loose” or worse, children born outside marriage were termed “illegitimate,” and couples who cohabited were said to be “living in sin.” In most circles, these terms have now been abandoned and even condemned—a manifestation of how much attitudes have changed.
Attitudes toward Sex, Marriage, and Childbearing
Consequences
Many researchers have attempted to examine the effects of single parenthood on children. The major challenge has been how to separate the effects of single parenthood from all of the other factors associated with being a single parent, such as being less educated, younger, or from a disadvantaged background or neighborhood as compared to being part of a married couple. Put differently, if it is these disadvantages that lead to single parenthood in the first place, then we could be wrongly attributing deficits in child development to family structure when it is really the result of a whole set of disadvantages associated with being a single parent.
Nevertheless, there are by now countless studies that have attempted to adjust for such differences, and they still find some negative effects for children, at least on average. These negative effects include dropping out of high school or college, a higher rate of crime and teen pregnancy, more behavioral or mental health problems, and lower levels of employment as an adult. 6 As Sara McLanahan and Christopher Jencks (2015) have written: “when taken together these studies are beginning to tell a consistent story. A recent review of 45 studies using quasi-experimental methods concluded that growing up apart from one’s father does reduce a child’s life chances in many domains.”
Changes in the family have also had economic consequences, such as higher rates of poverty, more income inequality, and less social mobility. Again, there is a chicken and egg problem here, but strong research clearly implicates family structure as one reason for each of these economic outcomes.
For example, McLanahan and Percheski (2008) find that family structure has become an important mechanism for the reproduction of class, race, and gender inequalities. In an analysis of trends between 1977 and 2012, Wilcox, Price, and Lerman (2015) find that state-level economic outcomes are significantly influenced by various measures of family structure even after controlling for a variety of other likely influences. States with more married parents have higher gross domestic product per capita, more mobility, lower child poverty, and higher median incomes.
In still another study, Sawhill uses a simple shift share analysis and then adjusts the results for selection bias to examine the impact of the growth of single-parent families on poverty rates. She finds that child poverty rates have risen by 5 percentage points or by 23 percent as a result of the growth of single-parent families between 1970 and 2013 (Sawhill 2014). 7 She argues that as social programs have reduced child poverty, family fragmentation has pushed rates in the opposite direction.
Finally, the breakdown of the family—by adding to the number of low-income households at the bottom of the distribution—has contributed to rising income inequality, explaining anywhere from 15 to 40 percent of increased inequality in recent decades (M. Martin 2006; Burtless 1999). 8
All of the above findings on the effects of family structure on children’s development and on economic outcomes need some interpretation. It is possible that what matters most is not just the structure of the family but its stability. As we noted, cohabitating relationships are usually short-lived, and just under half of these unmarried parents go on to form new relationships and have additional children with a new person or in some cases several new people. According to the Fragile Families study, 80 percent of the children born to unmarried parents experienced either family instability or their parent having a child with a new partner before they were age five (McLanahan 2011).
A second interpretation of the above findings is very simple. Two parents have twice as much time as one and can devote that extra time either to working or providing care to their children. It is worth remembering that a second income, even a low one, is likely to bring more resources into the household than the most generous social assistance program. This is one reason the poverty rate among single parents is four or five times as high as the rate among two-parent families (Moynihan 1996; Haskins 2015). Most of the latter have two paychecks, not just one. Some analysts control for income when looking at the effects of single parenting on children, but income is largely if not entirely endogenous (that is, a consequence of having two adults in the household). The same is true of parental time spent with children (Kalil, Ryan, and Chor 2014).
A third possibility is that individuals who make good partners also make good parents. They may simply have the maturity, the commitment, or the relationship skills that facilitate both. These kinds of personal attributes are not easily measured and controlled for in social science research and thus could explain some of the observed differences.
How one interprets this evidence is important. It does not mean that all single parents are doing a bad job or that all of the children growing up in such families are adversely affected. Nor does it mean that marriage is always the solution. Many of the fathers of these children are in prison, and are violent, addicts, or unable to play a positive role in their children’s lives (Edin and Nelson 2013; Wildeman and Western 2010). Finally, as emphasized in our discussion of the causes of the retreat from marriage, the relationship between changes in the economy, such as the disappearance of well-paid jobs for less skilled men, and the limited prospects of young women from low-income neighborhoods are part of this story. We cannot completely sort out the chicken from the egg and should recognize that more single parents means more poverty or related disadvantages—and more poverty and lack of opportunity, in turn, causes more single-parent families to form.
Although we have known for some time that children who grow up in single-parent families do not fare as well as those with two parents, some scholars are now finding that the consequences are especially serious for boys. Not only do boys need fathers, presumably to learn how to become men and how to control their impulses, but less obviously, and almost counterintuitively, it turns out that boys are more sensitive or less resilient than girls (Bertrand and Pan 2011). The kind of parenting they receive affects their development, including their behavior and their performance in school, more than parenting affects the development of girls.
Until now, these speculations about the disproportionate consequences of single parenthood for boys have been based on limited evidence. But new research from Stanford professor Raj Chetty and a team of colleagues (2016) shows that the effects on boys are clearly larger than for girls. They occur regardless of family income but are especially pronounced for boys living in high-poverty, largely minority neighborhoods. According to this research, when they become adults, boys from low-income, single-parent families are less likely to work, to earn a decent income, and to go to college not just compared with other boys but also compared with their sisters or girls who grew up in similar circumstances. These effects are largest when the families live in metropolitan areas (“commuting zones” 9 ) with a high fraction of black residents, high levels of racial and income segregation, and lots of single-parent families. In short, it is not just the boy’s own family situation that matters but also the kind of neighborhood in which he grows up. Exposure to high rates of crime and other potentially toxic peer influences without the constraining influence of adult males within their families seems to set these boys on a very different course than other boys, and perhaps more surprisingly, a different course from their sisters.
If single parenting affects boys more than girls, and an increasing number of boys are growing up in such families, the implications for the socialization of boys are troubling (Autor and Wasserman 2013). Moynihan was especially concerned about the lack of male authority in many inner-city families and communities. He (1965b) argued that “a community that allows a large number of young men to grow up in broken families, dominated by women, never acquiring any stable relationship to male authority, never acquiring any set of rational expectations about the future—that community asks for and gets chaos.” But Moynihan could not have predicted how, in combination with a criminal justice system that locked up a large proportion of young black men with little education, often for nonviolent offenses, so many children would be deprived of having a father in the home. In addition, when these men return to the community with a prison record, it becomes almost impossible for them to find work (Western and Wildeman 2009). Joblessness then starts the entire cycle over again, discouraging marriage, necessitating illicit activity, and potentially producing a new generation of boys without fathers.
Although the research to date has focused on boys, we should remember that most single-parent families begin with an unwed and unplanned pregnancy and that these are more likely to occur when a girl grows up in a single-parent family in a distressed neighborhood. So providing the girls in such communities with both the motivation (access to education, decent jobs, and alternative sources of identity beyond motherhood) and the means (access to the most effective forms of contraception) to avoid an early unwed birth needs to be part of the response as well.
Before we discuss evidence on ways to increase marriage rates and reduce the negative impacts of single parenting, though, we think it useful to provide a brief summary of what research on marriage suggests is the reason living in a married-couple family is best for children’s development. The extensive literature on marriage has found evidence from one or more studies that supports each of a large number of mechanisms that contribute to the relationship between a stable, two-parent family structure and child well-being. A recent masterful review by marriage scholar David Ribar (2015) reports evidence on a host of household and parent attributes that are important to improving child outcomes, including total household income, parents’ ability to “specialize” in certain aspects of household productivity, father involvement in childrearing, parents’ physical and mental health, parenting quality, social supports, health insurance, home ownership, and family stability. All of these favor married-couple families. Ribar argues that “the likely advantages of marriage for children’s wellbeing are hard to replicate through policy interventions other than those that bolster marriages themselves” (p. 23). Ribar’s conclusion is that, as in so many other issues of child development, there is no single mechanism that exhaustively explains any given child outcome. Explaining the advantages of marriage for child development requires appeal to a host of mechanisms.
What to Do
Increasing the share of children and adults in married-couple or stable cohabiting families would have three likely impacts (American Enterprise Institute and Brookings 2015). Given the very high poverty rates that characterize mother-headed families, the first impact would be a decline in poverty rates, especially for children. A decline in poverty rates would also necessarily mean a decline in income inequality. Second, child development would improve because children in families headed by women experience more problems than children in married-couple families, as we noted. Third, increased marriage rates would likely contribute to the well-being of adults, especially men (Waite and Gallagher 2000).
But the policy solutions for which we have evidence suggest that even well-financed and implemented policies would not reverse the trends in family composition or fully ameliorate their consequences. On the other hand, we do have evidence that some policies produce modest impacts and others might be called promising.
We consider five types of policies or programs that could increase marriage rates or the stability of cohabitation. These include policies to reduce nonmarital births, make tax and transfer policies more profamily, improve couple relationships, use media campaigns to emphasize the importance of family stability for children, and help young men to become better partners. Even if these policies were successful, we would still have millions of children being reared in female-headed families who need support, so we should also help them to avoid poverty and achieve a reasonable level of financial well-being.
Birth control to reduce unplanned pregnancies and nonmarital births
More or better use of birth control could reduce teen pregnancy rates, unintended pregnancies at older ages, and abortion rates. In addition, by reducing the number of single-parent families, it could reduce poverty and income inequality and promote children’s development. On top of all of these benefits, birth control saves the government money. In fact, it already produces this entire range of benefits, but more effective use of birth control would expand them.
In recent years, there have been several large-scale studies of efforts to increase the voluntary use of birth control, especially the use of long-acting, reversible contraception (LARC) by low-income mothers. These prospective studies involved making contraception free, training medical personnel, conducting social marketing campaigns to encourage the use of birth control, and using better counseling to explain the advantages and disadvantages of various types of contraceptives.
One of these studies, conducted in the St. Louis area and called the Contraceptive CHOICE Project, enrolled more than 9,000 low-income mothers and gave them the option of choosing their own method of birth control at no cost (Winner et al. 2012). Participants were 14 to 45 years old; were either not using any contraception or were willing to consider switching to a different method; did not want to become pregnant for at least the next 12 months; and were either sexually active or planning to be sexually active with a male partner during the next six months. At the end of three years, the pregnancy rate for those who used the pill, patch, or ring was 9.4 % the rate for those who used LARCs (IUDs and implants) was 0.9 % and the rate for those who received injections was 0.7 percent. Similar large-scale studies have been conducted in Iowa and Colorado with similar results (Biggs et al. 2015; Ricketts, Klingler, and Schwalberg 2014).
None of these studies used random assignment. Although the size of the effects (compared with what was happening in other cities or states) makes it unlikely that such effects are due to other factors operating at the same time, we cannot know this for sure. Another study involving fifteen states, led by the Bixby Center at the University of California, San Francisco, did involve a randomized control trial of forty clinics (Harper et al. 2015). The Bixby study found that provider training alone (with no reduction in the cost of birth control) caused the unplanned pregnancy rate to drop by almost half in one year’s time for women who came to the clinic seeking family planning. Another group of women in the study came to the clinics seeking an abortion and were offered a LARC afterward. In these cases, there were no significant differences between the pregnancy rates of women in the treatment and control clinics. The authors believe the lack of effect in these abortion-related cases was primarily because women who wanted a LARC were not easily able to obtain one due to greater cost barriers (the cost is less likely to be covered by a government program when a woman gets an abortion) and the need to return to the clinic for a second visit to obtain the LARC.
There have also been effective efforts to reduce teen pregnancy rates. Since the early 1990s, teen birth rates have declined almost every year and have fallen by over 50 percent, from 59.9 per 1,000 teen females in 1990 to 26.5 per 1,000 in 2013 (J. Martin et al. 2015). Even so, American teenagers still have much higher birth rates than teens in many other nations with advanced economies. Japan, Denmark, and the Netherlands, for example, all have rates under 5 per 1,000 (Ventura, Hamilton, and Matthews 2014).
To further drive down the teen pregnancy rates, the Obama administration has launched a major initiative. Working with the research firm Mathematica Policy Research and Child Trends, the administration reviewed all published and unpublished evaluations of teen pregnancy prevention programs that they could find (Mathematica Policy Research and Child Trends 2010). After reviewing nearly 1,000 studies in accord with detailed procedures developed by the Department of Health and Human Services, the team identified thirty-one model teen pregnancy prevention programs (more were identified later) with rigorous evidence of impacts on sexual activity, use of contraceptives, sexually transmitted infections, or pregnancy rates. The administration is now funding and evaluating 75 initiatives that replicate one of these model programs, enrolling more than 100,000 teens annually in 37 states. The administration hopes to identify the most effective programs and encourage expansion of those programs.
One of the most impressive findings from research on family planning is the number of studies that have shown net savings from subsidized payments for birth control, as a recent review on the website the Incidental Economist demonstrates (Liebman 2014). Here’s the basic math: the average insurance payment for a vaginal delivery is around $18,000; the cost of a C-section is $28,000. If there are complications, these costs can skyrocket. By comparison, the average cost of contraception is between $100 and $600 annually (Klein 2014). Combine these numbers on cost data with the facts that nearly 70 percent of births to unmarried women ages 20–29 are unplanned and that, when given a choice between types of birth control provided without charge, as many as 70 percent of low-income women select the most effective form of birth control (LARCs), it is clear that there are serious savings to be had by expanding the availability of subsidized birth control to low-income women. Although these women may go on to have children later, they often gain more education, get experience in the labor market, and form more stable relationships by delaying parenthood. At least four studies have produced estimates of the benefit-cost ratios of expanded use of effective contraception; the estimates range from savings of $3.74 to $7.00 for every $1 spent on birth control (Sawhill, Thomas, and Monea 2010; Frost, Finer, and Tapales 2008; Frost, Henshaw, and Sonfield 2010; Thomas 2012).
There is little doubt that birth control initiatives, especially the increased use of LARCs, will reduce nonmarital and unplanned births. Reductions in these births will in turn save money. In addition, avoiding nonmarital births at an early age can increase the chance that women will marry later in life (Lichter and Graefe 2001).
Reduce marriage penalties in tax and transfer programs
The tax code and means-tested programs can present disincentives for marriage, because single people who marry and combine their incomes could see higher taxes and fewer means-tested benefits. The earned income tax credit (EITC), which is intended primarily to increase work incentive and augment the income of low-income workers, especially parents, is designed so that qualified workers receive more money as they earn more up to a certain amount; then their EITC payment is flat for several thousand dollars of additional earnings; then the EITC payment phases out over a broad income range. But if a mother with two children and $18,000 in earnings marries a man earning $32,000, her EITC falls from the maximum of $5,548 to zero. The amount of the EITC marriage penalty or incentive hangs on the combined income of the husband and wife.
One of the few studies to estimate the impact of marriage on the size of the EITC and cash welfare based on a sample representative of the population was conducted by Greg Acs and Elain Maag (2005) 10 of the Urban Institute. The authors identified 744 cohabiting couples with children in the Survey of American Families who had a combined income under 200 percent of poverty. They calculated the impact that marriage would have on their EITC benefit as it existed in 2008 as well as the couples’ Temporary Assistance for Needy Families (TANF) welfare benefit, if the mother received one.
A major finding was that 75 percent of the cohabiting low-income couples would receive a marriage bonus from the EITC, while only 10 percent would receive a penalty (the remaining 15 percent would experience little to no change). The average increase in the EITC for the 75 percent who received it would be about $1,400. Other tax code exemptions, deductions, and credits that these couples could qualify for if married increased the marriage bonus to a total of around $2,400. For the 10 percent who were hit with a marriage penalty from the EITC, the average total penalty was around $1,750.
Turning to the TANF program, because TANF benefits phase out rapidly as earnings increase, almost all the cohabiting couples who received TANF would have their benefit reduced if they married. But only 14 percent of the couples were receiving TANF benefits. For this small minority of couples, the reduction in the TANF benefit was between $1,800 and $2,100. Of the 14 percent of couples who received TANF, fewer than 4 percent got both a tax penalty and a TANF reduction; for these families, the combined loss was substantial, about $3,300. But 70 percent of the 14 percent who received a TANF reduction also received an EITC bonus. The combined tax bonus and TANF reduction for these couples still left them with a net marriage bonus that averaged $1,300.
Two conclusions are justified. First, a small minority of cohabiting couples with combined income under 200 percent of poverty who marry would be subjected to an EITC marriage penalty. Second, the marriage penalty for the relatively small group of mothers and fathers who receive TANF seems likely to be substantial. Moreover, the study considered only the EITC and related tax credits and the TANF cash benefit, but other welfare benefits such as Medicaid, food stamps, housing, school lunch, and child care also have phase-out rules. A recent study by researchers at the Urban Institute found that nearly 80 percent of a representative sample of families with children with incomes below 200 percent of the poverty line received at least one welfare benefit, and 45 percent received two or more (Edelstein, Pergamit, and Ratcliffe 2014). In many cases, there would be marriage penalties from these programs. An especially serious disincentive occurs in the Medicaid program, where eligibility ends abruptly at a given income level, which varies widely across states (Centers for Medicare and Medicaid Services 2014; Garfield and Damico 2014).
It follows from these considerations about means-tested benefits that we should worry more about the marriage penalty that low-income couples encounter from means-tested programs than about the EITC and other tax credits, especially because the Tax Relief Act of 2010 extended the bottom 15 percent tax bracket for married couples filing jointly, increased the standard deduction, and extended the EITC phase-out range for married couples, all of which reduce the tax penalty associated with marriage.
Extending the phase-out range for tax benefits to reduce the disincentive to work for married couples would be expensive. As one moves up the income scale into the range where many married couples are likely to be located, many households become eligible for the reduced tax bill, leading to big revenue losses. In addition, there is not good evidence that these changes in taxes or benefits have strong effects on people’s propensity to marry. Thus, the benefit-cost ratio of such policies is likely to be small.
Improving the relationships of married and cohabiting couples
Based on the view that marriage and even improved relationships among unmarried couples would be good for the adults and children involved, the George W. Bush administration launched a marriage initiative in 2001 to determine whether marriage education and associated services for couples might improve relationship quality and help couples either to get married or to prolong their relationship and, if so, whether these impacts would in turn have a positive impact on children’s development and behavior.
A major part of the Bush initiative was the Building Strong Families (BSF) program, evaluated by the research firm Mathematica. The program aimed to strengthen the relationships and parenting skills of young couples who had a baby together outside of marriage. The program was implemented with more than 5,000 couples randomly assigned to an experimental or a control group at eight sites. Parents in the experimental group were offered marriage education classes in groups, using a formal curriculum, as well as advice and support from a family-services coordinator.
The evaluation measured the quality of the couples’ relationships, their coparenting relationships, family stability, children’s social-emotional development, and other outcomes. These measures were collected both at 15 months and 36 months after participants had enrolled in the program. At 15 months, the BSF program produced few significant effects, including no effects on whether the couples stayed together or got married (Wood et al. 2010). However, the Oklahoma program produced a pattern of positive effects, while the Baltimore program saw some negative impacts, including a slight increase in physical assaults by the father. The positive effects in Oklahoma included relationship happiness, parenting skills, support and affection, use of constructive behaviors to resolve conflicts, avoidance of destructive conflict behaviors, marital fidelity, quality of coparenting, whether the father lived with the child, and whether the father provided substantial financial support.
Mathematica’s 36-month follow-up again showed few impacts across the eight sites (Wood et al. 2012). Although most of the Oklahoma impacts had disappeared by 36 months, one important remaining difference was that children in the treatment group were about 20 percent more likely to still be living with both of their parents as compared with children in the control group.
The second Bush marriage initiative, called Supporting Healthy Marriage, was similar to BSF, but attempted to increase the relationship skills of couples who were already married rather than unmarried couples. Results from Supporting Healthy Marriage are somewhat more encouraging than those obtained from BSF, but most of the effects were not statistically significant, and even the significant effects were modest. More importantly, program couples were no more likely to stay together, and there were no effects on measures of their children’s behavior or development (Lundquist et al. 2014).
The Bush administration initiative was the first large-scale effort to develop marriage programs for poor couples and to test their effectiveness. The fact that the initial effort to conduct such large and complex programs produced disappointing results should not be too surprising. It is not inconceivable that the programs could be improved over time. This is especially the case because other high-quality studies have shown that marriage education can have a positive effect on couples’ relationships and breakup rates, although these studies did not focus exclusively on low-income couples (Hawkins 2013; Schulz, Cowan, and Cowan 2006; Stanley et al. 2010).
Moreover, Philip and Carolyn Cowan, two of the most experienced researchers and designers of couple relationship programs, recently reviewed the evidence on education programs for couples and reached three conclusions: first, without intervention, “average couple relationship satisfaction declines”; second, including fathers in the programs “results in value-added contributions to family functioning”; and third, eight of nine studies of couple relationship programs that include child outcomes show benefits for children (Cowan and Cowan 2014). That only two of the nine studies reviewed by the Cowans focused on low-income families somewhat limits the application of their conclusions to low-income families.
Thus, it may be premature to abandon Bush’s marriage education initiatives. It would be especially useful to replicate the Oklahoma program with a focus on finding ways to reduce its costs and maintain its impacts. Another issue that needs more study is participation rates. Only 61 percent of those who enrolled in the programs attended even one session, and among those who did attend, the median curriculum completion level was only 40 percent (Dion et al. 2008). It seems unlikely that any curriculum can be effective when so many participants never show up and a majority of the rest miss more than half the sessions.
Help young men
In his heralded 1987 book The Truly Disadvantaged, sociologist William Julius Wilson argued that unemployment among young black men was a key to explaining the decline of marriage among black couples. Wilson constructed a “black marriageable male index” based on comparing the number of employed black men to the number of black women in the same age range. He shows that in 1960, the ratio was about 70 employed black men for every 100 black women in the 20 to 24 age range. By the 1980s, the ratio had fallen to 50 employed black men for every 100 black women.
In addition to their high rates of unemployment, nearly 60 percent of black high school dropouts born between 1965 and 1969 had been in prison by the time they reached their early thirties (Pettit and Western 2004). Having a prison record makes it difficult to find work when men leave prison. Prison also disrupts their relationships with relatives—especially children—and friends. It would be hard to imagine a combination of factors that would do more to reduce marriage prospects than a lousy work history and a prison record.
One reason young men may have such difficulty with the law is that they have little or no consistent contact with their fathers. David Autor and Melanie Wasserman (2013) review evidence showing that single mothers spend less time with sons and harshly discipline them more often than daughters. Similarly, they note that although boys in general act out in school more often than girls, the gap is greater for boys and girls from female-headed families than for boys and girls from married-couple families. Boys also see their fathers much less after their parents separate, and given the high rates of single parenthood in low-income and minority communities, the bond between fathers and sons is more often disrupted in these communities (Carlson and Turner 2010).
At least two public policies could improve young men’s chances of finding work and help them to develop healthy relationships with young women. The two policies are support for programs that prepare them for employment (Mead 2011) and that reduce their rates of incarceration. Among the former, several programs that have been tested by random-assignment evaluations have shown positive impacts on young men’s employment (Holzer 2014). Foremost among them are the Career Academies program and apprenticeship programs that give young people a skill and a certificate, often through community colleges (Kemple and Willner 2008; Lerman 2014). The Career Academies program even led to higher marriage rates.
But Career Academies may be an outlier. Daniel Schneider (2015) reviewed sixteen experimental programs involving early childhood development, workforce training, and income support that aimed to improve the economic well-being of low-income men and women. Most of the programs produced positive effects on the economic well-being of young men, young women, or both, but only a few, including Career Academies, had strong impacts on marriage rates. Based on Schneider’s review, there is only modest evidence that programs that increase economic well-being also increase marriage rates.
There is a growing consensus that states and the federal government should loosen mandatory sentencing laws and thereby reduce the number of nonviolent offenders who serve long prison sentences. Many states, sometimes forced by budget shortages, are already beginning to change their mandatory sentencing laws, although we know little about the effects of these changes. At the federal level, many politicians from both parties have proposed reforms of mandatory sentencing laws for nonviolent offenses as well as new or improved prison release programs to help former prisoners adapt to civilian life, especially by finding a job (Jalonick 2016; Wildeman, Wakefield, and Lee 2016).
In addition to Career Academies and apprenticeship programs, both evaluated by large randomized evaluations, another program for young men that shows promise based on a large randomized trial is the Becoming a Man (BAM) program developed by the Crime Lab at the University of Chicago (2012). Reflecting the principles of cognitive behavioral therapy, BAM enrolled 2,740 male students in grades 7 through 10 from 18 Chicago schools and randomly assigned approximately half of them to receive the program. Students selected for the program were considered to be at risk for behavioral problems but, based on their school records, were nonetheless likely to attend school enough to benefit from the BAM intervention. They participated in the program during school hours, during after-school hours, or both. The during-school program consisted of twenty-seven one-hour small-group sessions that met weekly during the school year. Each session aimed to develop a specific cognitive-behavioral skill such as self-regulation or interpersonal problem solving. The after-school program consisted of sports activities and emphasized conflict resolution skills and social and emotional learning.
Using school records, a composite measure was formed consisting of number of days present, grade point average, and end-of-year enrollment. Data on criminal behavior was obtained from records maintained by the Illinois Criminal Justice Information Authority. The composite school engagement outcome showed significant differences during both the program year and the next school year between program and control students, favoring students participating in the program. Similarly, arrest records showed that program students had 44 percent fewer arrests for violent crime than control students. The researchers concluded that these impacts “provide the most rigorous, large-scale evidence to date that a social-cognitive skill intervention can improve both schooling and delinquency outcomes for disadvantaged youth” (University of Chicago Crime Lab 2012, 4).
Although examination of national data on marriage, school performance, employment and earnings, and arrests does not provide much room for optimism that the nation is making progress with young, minority males, programs such as Career Academies, apprenticeships, and BAM show that knowledge of how to improve the lives of these vulnerable young men is increasing. These and similar programs should be expanded; meanwhile, program development and evaluation must continue to discover new programs that can improve the life chances of young men.
Help single mothers
We may hope that the decline in marriage rates and increase in nonmarital birth rates will turn around, but meanwhile a huge share of the nation’s children will continue to live in female-headed families. For this reason, maintaining or even expanding the focus of state and federal policies on these female-headed families is warranted.
The nation has taken two broad approaches to help poor single mothers and their children. One is to provide cash and noncash support. The federal and state governments now spend about $1 trillion annually on these programs, a considerable portion of which goes to female-headed families (Congressional Research Service 2012). 11 The major programs included in this estimate are Medicaid, food and nutrition programs, Supplemental Security Income, the EITC, the Additional Child Tax Credit, and housing programs. The second approach is to encourage poor mothers to work, usually at low-wage jobs, and then use government programs to subsidize their earnings. One of the great tensions in American social policy centers is whether it is better to give welfare benefits to able-bodied mothers or to encourage and cajole them to work and then subsidize their earnings, which are often below the poverty level (Mead 1986; Haskins 2006). A key event in the work approach was passage of the 1996 welfare reform law, which greatly strengthened work requirements and gave states an incentive to enforce them.
Although the welfare reform law had shortcomings, its passage was followed by a substantial increase in the proportion of poor single mothers who were employed. We can get a good idea of the course of the employment-to-population ratios after welfare reform by comparing the ratio for the average of the five years before welfare reform (1991–95) with the average ratio of the five years following welfare reform (1997–2001). The former average was 46.4 % the latter was 62.6 percent—an increase of 35 percent. This may well be the biggest increase in work rates over a short period for any demographic group in American history. The recessions of 2001 and 2007–9 reduced the employment-to-population ratios of never-married mothers (and nearly every other group). For never-married mothers, the decline was from a high of 66 percent in 2000 to a low of 57 percent in 2011. But as the economy recovered from the Great Recession of 2007–9, the employment-to-population ratio began to move up again in 2012. The 60 percent rate for never-married mothers in 2014 is nearly 30 percent above the comparable rate in the five years before welfare reform. A careful analysis of the composition of income among the nation’s mother-headed families with children by the nonpartisan Congressional Research Service shows that today federal programs reduce the poverty rate among these families by about 50 percent (Gabe 2014; Heinrich and Scholz 2011).
However, during the recessions that began in 2001 and 2007, work rates fell and poverty rates increased among mother-headed families, showing that, like other families, single-mother families depend on the economy to generate jobs if they are to continue making economic progress. Another piece of bad news is that some mothers have problems making the transition to work and often use up their time-limited TANF benefits, are eliminated from the rolls for rule violation, or leave the rolls voluntarily, perhaps to work at a job they later lose. This group of mothers lacks both earnings and TANF benefits. In one study, their annual income was $6,178, compared with $17,681 for working mothers who left TANF. Not surprisingly, these mothers and their children have high rates of poverty (Blank and Kovak 2007; Loprest and Nichols 2011).
Several policy changes could help poor, single working mothers increase their income and in some cases escape poverty. First, states could do a better job of helping the mothers prepare for and find work (Germanis 2015a, 2015b). Second, we could do more to ensure that these mothers and children get child support, especially by persuading states, perhaps with financial incentives, to give all child support collections to the mothers by ending the state and federal practice of retaining part of child support payments as an offset to welfare or other benefits. A third reform would be to help states mount work programs for noncustodial fathers who owe child support so that they have earnings with which to make payments. Other worthwhile improvements in the work support system would be to expand child care subsidies and to provide subsidized jobs for mothers and fathers who cannot find work. Helping more low-income parents with their child care bill would increase their incentive to work, provide an income supplement, and reduce a serious inequity in current law that allows only some low-income working families to receive a child care subsidy while similar families receive no subsidy. During the Great Recession, states used federal emergency funds to subsidize around 260,000 jobs, most of them in the private sector (Pavetti, Schott, and Lower-Basch 2011). Some analysts argue that the federal government should make subsidized employment a permanent feature of federal policy (Dutta-Gupta et al. 2016). Developing state expertise in subsidizing jobs would be especially appropriate if Congress strengthened the work requirements in the nation’s food stamp and housing programs to extend the message that the able-bodied must work or prepare for work as a condition of receiving means-tested benefits.
Concluding Thoughts
The breakdown of the family has increased the nation’s poverty rate, increased income inequality, adversely affected the development of millions of children, and increased spending on social programs. Some scholars believe that for these reasons we should do more to encourage marriage and not give up just because some programs have had disappointing effects so far (Haskins 2014, 2015; Wilcox and Wolfinger 2016). Other scholars question whether restoring marriage to its former status as the central feature of American family life and the culturally accepted way to raise children is possible, even if desirable (Sawhill 2014). Reversal of demographic trends that have been moving in the same direction for four decades seems unlikely, especially in low-income communities of color where marriage has virtually disappeared. In this context, it makes sense to both discourage unplanned childbearing outside of marriage and continue to provide assistance to single-parent families, ensuring that single mothers with little education gain at least a modicum of financial security. Of course, we could work simultaneously on increasing marriage rates and family stability, reducing nonmarital and unplanned births, and helping low-income single mothers care for their children. We need to do all three if one believes, as we do, that children who live with two parents and are wanted and cared for will have better lives as a result.
Footnotes
NOTE:
Ron Haskins and Isabel Sawhill were the winners of the 2016 Daniel Patrick Moynihan Prize, awarded annually by the American Academy of Political and Social Science to recognize individuals who are champions of social science in the public realm—women and men whose careers demonstrate how public policy can be more effective when it is informed by sound science. Each year, we invite the Moynihan Prize recipients to deliver a public lecture on a topic of their choosing. This is the third year that the lecture has been published as a stand-alone piece in The ANNALS.
Notes
Ron Haskins holds the Cabot Family Chair in Economic Studies at the Brookings Institution, where he codirects the Center on Children and Families. He is also a senior consultant at the Annie E. Casey Foundation. He is the president of the Association for Public Policy Analysis and Management.
Isabel V. Sawhill is a senior fellow in Economic Studies at the Brookings Institution, where she has also been a vice president and center director. She served in the Bill Clinton administration as an associate director of the Office of Management and Budget.